JP2022014409A - Machine tool and control method therefor - Google Patents

Abstract

To remove chips more efficiently.SOLUTION: A machine tool, which removes chips more efficiently, comprises: a fluid injection part comprising a nozzle which can control a fluid injection direction; an imaging part which images the interior of the machine tool; a chip detection part which detects a chip position in the machine tool on the basis of a picked-up image acquired from the imaging part; a derivation part which derives a correction amount of a fluid injection target position from the chip position according to the chip position; and a control part which controls the fluid injection direction by the nozzle by use of the correction amount.SELECTED DRAWING: Figure 1

Description

The present invention relates to a machine tool and a control method thereof.

In the above technical field, Patent Document 1 discloses a technique for removing chips in a machine tool with a robot including a manipulator. In particular, in paragraph 0059, "The control unit 12 determines the position and orientation of the nozzle 18. Here, in the present embodiment, the first area 108a, the second area 108b, the third area 108c, And the position and orientation of the nozzle 18 (ie, the tool coordinate system) for effectively removing the machining debris present in the fourth area 108d is predetermined by the user. " Further, in paragraph 0060, "As an example, the tool coordinate system of the nozzle 18 is set so that the work chips existing in the region 108 can be blown outward radially from the center of the jig 104. Will be done. "

Japanese Unexamined Patent Publication No. 2017-35765

However, the above-mentioned document determines from which direction the fluid is sprayed on the work chips at which position, but as described in paragraph 0037, it is said that "the fluid is blown onto the work chips to remove the work chips". There was only an idea, and it was not possible to effectively remove the chips.

An object of the present invention is to provide a technique for solving the above-mentioned problems.

The present invention provides machine tools and the like described in the claims.

According to the present invention, chips can be removed more efficiently.

It is a figure which shows an example of the structure of the machine tool and the coolant control which concerns on 1st Embodiment. It is a figure which shows an example of the structure of the machine tool and the coolant control which concerns on 1st Embodiment. It is a figure which shows an example of the structure of the machine tool and the coolant control which concerns on 1st Embodiment. It is a figure which shows an example of the structure of the machine tool and the coolant control which concerns on 1st Embodiment. It is a figure which shows the structure of the chip detection part of the machine tool which concerns on 1st Embodiment. It is a figure which shows the learning method of the chip detection part of the machine tool which concerns on 1st Embodiment.

Hereinafter, embodiments of the present invention will be described in detail exemplary with reference to the drawings. However, the components described in the following embodiments are merely examples, and the technical scope of the present invention is not limited to them.

[First Embodiment]
The machine tool 100 as the first embodiment of the present invention will be described with reference to FIG. The machine tool 100 is a machine for cutting or grinding a work 130 made of metal, wood, stone, resin or the like by using a spindle and a tool (not shown). As shown in FIG. 1, the machine tool 100 includes a coolant injection unit 101, a camera 102, a chip detection unit 103, a lead-out unit 104, and a control unit 105. The machine tool 100 also includes a tool spindle 151, a holder 152, and a tool 153.

The coolant injection section 101 injects the coolant 112 not only for lubrication, cooling, and cleaning of the spindle, tools, and work 130, but also for effective discharge of chips 110. The coolant injection unit 101 has a coolant nozzle 111 capable of controlling the method of injecting the coolant. The injection method includes at least the direction of the coolant nozzle 111 (or the injection target position 142 that defines the direction), the injection pressure of the coolant 112, the diffusion level of the coolant, how to move the coolant nozzle 111, and the presence or absence of pulsating current of the coolant 112. Includes any one. The diffusion level is defined by the shape or area of the area reached by the coolant 112. The way of moving the coolant nozzle 111 is the locus of the injected coolant 112. Here, the fluid to be injected for discharging the chips is exemplified as a coolant, but any fluid such as air or water can be applied as long as the fluid can move the chips.

The camera 102 is an image pickup unit that acquires an in-machine image by taking an image of the inside of the machine tool 100. The camera 102 may be a camera fixed near the ceiling in the machine, or may be a camera that can be attached to and detached from the spindle like a tool by an automatic tool changer (ATC). Further, the camera 102 may be configured to be slidable along a rail (not shown) attached to the ceiling inside the aircraft. Here, the camera 102 is provided as an example of the image pickup apparatus, but for example, a line scanner, an X-ray photographing apparatus, or the like may be provided.

The chip detection unit 103 detects the position 141 of the chips 110 as a cleaning object existing inside the machine tool 100 based on the in-machine image captured by the camera 102. Specifically, the in-flight image is divided into small areas (meshed), and the presence or absence of chips 110 is determined for each partial area. In addition to the position of the chip 110, the chip detection unit 103 may detect various information about the chip, such as the size of the chip, the amount of the chip, the shape of the chip, the type of the chip, or a combination thereof. For example, a "class" indicating the ease of cleaning chips may be determined.

The chip detection unit 103 extracts the feature amount of the captured image. The feature amount extracted by the chip detection unit 103 includes at least one of the information amount, the frequency component, the contrast, and the luminance distribution of the captured image. The chip detection unit 103 extracts the feature amount of the captured image and determines whether or not the image can be used for chip detection.

The derivation unit 104 uses the chip position using the injection method model 107 generated by machine learning the chip position, the injection method of the fluid injected in the past, and the change in the chip position after the fluid injection as teacher data. The injection method for injecting the fluid, which is predicted from, is derived. The lead-out unit 104 may, for example, derive the correction amount C from the chip position 141 of the coolant injection target position 142 according to the chip position 141. After being ejected from the nozzle 111, the coolant 112 draws a parabola and reaches the arrival position 143. In order to move the chips 110 towards the shooter 120, the coolant 112 should arrive at the arrival position 143 on the left side of the figure of the chips 110 to generate a flow to the shooter 120, rather than hitting the chips 110 directly. Is. Therefore, the injection target position 142 is determined in consideration of the behavior of the coolant 112 after such ejection and the influence on the chips after the coolant arrives. The lead-out unit 104 may derive the injection method according to not only the chip position 141 but also the chip amount. For example, if the amount of chips is large, the coolant 112 may be first directly hit against the chip positions to disperse the chips, and then the injection target position 142 may be set at a position deviated from the chip position 141.

The injection method model 107 was generated by machine learning using the injection method of the coolant 112 injected in the past, the horizontal distance between the coolant nozzle 111 and the chip position 141, and the change in the chip position after fluid injection as training data. It is a model. The change in the chip position 141 after the fluid injection is derived by comparing the captured images before and after the injection of the coolant 112. That is, it is determined how the position of the target chip changes before and after the injection of the coolant 112, and if the amount of change exceeds a predetermined value and the change direction is in the direction of the shooter 120, a high score is obtained. Will be.

The coolant injection unit 101 may include two or more nozzles 111, and the extraction unit 104 may derive the optimum direction of each of the two or more nozzles for each chip position 141 as an injection method. ..

The control unit 105 controls the imaging of the camera 102 in conjunction with the movement of the coolant pump 106 while controlling the coolant pump 106. The control unit 105 controls the coolant pump 106 to stop the injection of the coolant 112, then controls the camera 102 to perform an image pickup, and controls the coolant pump 106 to inject the coolant again after the image pickup.

Further, the control unit 105 determines whether or not it is the timing to perform photographing for chip detection according to the machining program (G code) or the user operation. For example, a break in machining during running (timing to scrape another surface of the work), timing to replace a tool by ATC, and timing instructed by a user correspond to this.

The control unit 105 controls the coolant injection unit 101 based on the result of analyzing the captured image of the camera 102, and effectively and efficiently injects the coolant into the chips.

The derivation unit 104 may derive the correction amount C according to the chip position 141 without using the injection method model 107. For example, the derivation unit 104 updates the storage unit that stores the chip position 141 and the correction amount C in association with each other, and the correction amount C stored in the storage unit according to the change of the chip position 141 after the coolant injection. It may have an update unit. Here, the correction amount C is the difference between the injection target position 142 and the chip position 141 when the position where the vector indicating the direction of the coolant nozzle 111 collides with the bottom surface of the machine tool is the injection target position 142, and is the plane distance. Is. The injection target position 142 and the chip position 141 can be defined by positions on the XZ plane on the bottom surface of the machine tool 100. However, the definition of the correction amount C is not limited to this, and may be, for example, an angle formed by the direction of the coolant nozzle 111 and the direction connecting the position of the coolant nozzle 111 and the chip position. The lead-out unit 104 may lead out the correction amount C according to the distance between the position of the coolant nozzle 111 and the chip position 141. In the case of a configuration in which the coolant nozzle 111 itself can be moved, the lead-out unit 104 may derive the correction amount C in consideration of the moving speed of the coolant nozzle 111.

The lead-out unit 104 may further derive the correction amount C according to the injection pressure of the coolant 112 by the coolant injection unit 101. For example, if the injection pressure of the coolant 112 is high, the difference between the injection target position 142 and the arrival position 143 of the coolant becomes small, so the correction amount C may be derived so as to be small.

Further, as shown in FIG. 2, when the chip position 141 is close to the wall 201, the outlet unit 104 is ejected toward the wall 201 so that the reflected coolant 112 moves the chip 110 to the shooter 120. , The correction amount C is derived. Similarly, as shown in FIG. 3, when the chips 110 are stuck in the corners 301 of the machine tool 100, the coolant 112 is ejected toward the corners 301, and the reflected coolant 112 moves the chips 110 to the shooter 120. The derivation unit 104 derives the correction amount C so as to cause the correction.

The lead-out unit 104 may further derive the injection pressure of the coolant 112 based on the chip position 141. Specifically, if the chip position 141 is far from the coolant nozzle 111, a strong injection pressure may be derived, and if it is close, a weak injection pressure may be derived. The control unit 105 controls the coolant injection unit 101 so as to inject the coolant 112 with the derived injection pressure.

The lead-out unit 104 may further derive the diffusion level and the way of moving the coolant 112 according to the chip position 141. For example, as shown in the left figure of FIG. 4, in the lead-out unit 104 and the control unit 105, when the chip positions 141 are spread and distributed along the wall 401, the arrival positions of the coolant 112 are zigzag along the wall 401. The coolant nozzle 111 may be moved to reciprocate to control the chips 110 away from the wall. Further, for example, as shown in the right figure of FIG. 4, in the lead unit 104 and the control unit 105, when the chip positions 141 are spread and distributed along the wall 401, the coolant 112 spreads widely along the wall 401. The chips 110 may be controlled to move away from the wall by injecting the chips 110 in such a manner.

The control unit 105 controls the imaging of the camera 102 in conjunction with the movement of the coolant pump 106 while controlling the coolant pump 106. The control unit 105 controls the coolant pump 106 to stop the injection of the coolant 112, then controls the camera 102 to perform an image pickup, and controls the coolant pump 106 to inject the coolant again after the image pickup.

For example, the control unit 105 turns off the coolant pump 106 at the timing of tool replacement by the ATC (Auto Tool Changer) to stop the injection of the coolant 112, and photographs the chips according to the stop time.

FIG. 5 is a block diagram for explaining the internal configuration of the chip detection unit 103. The chip detection unit 103 includes a calculation processing unit 501, a storage 502, and a touch panel 503.

The arithmetic processing unit 501 executes various arithmetic processing and realizes various functions. The storage 502 stores various data and functions as a working area when the arithmetic processing unit 501 performs arithmetic processing. Specifically, the storage 502 stores the chip detection program 521, the determination parameter 523, the determination result 524, the cleaning condition 525, the number of cleanings 526, and the in-flight image 527. Here, the storage 502 is provided inside the chip detection unit 103, but various data may be provided to the chip detection unit 103 via the network from the storage prepared in the external server.

The touch panel 503 has an input function for receiving an instruction input from a user and a display function for displaying a chip determination result or the like.

The chip detection program 521 adopts a chip determination model 522 by applying the determination parameter 523 to the learning / inference model. The determination parameter 523 is a parameter for detecting chips, and is obtained by pre-learning a learning / inference model using information about a partial image and chips as teacher data.

The determination parameter 523 is stored for each of a plurality of different mesh sizes. Further, the storage 502 separately stores learning parameters (not shown) that influence the characteristics such as the learning efficiency of the chip determination model 522.

The determination result 524 is the result of executing the chip detection program for each of the partial images (corresponding to the mesh area) obtained by dividing the in-flight image 527 into meshes.

As shown in FIG. 6, the chip determination model 522 derives a "class" indicating the ease of cleaning the chips for each partial image 601 as the determination result 524. This "class" is determined by comprehensively considering the amount of chips, the degree of density, size, length, shape, and the like. Specifically, the lower the class is, the easier it is to clean, such as less chips, scattered, small, short, and hard to get caught. The more difficult it is to clean, such as easy, the higher the class. For each class, the probability of that class is stored as the determination result 524. For example, if three classes of "class 0" with no chips, "class 1" with few chips, and "class 2" with many chips are set, the probability of being "class 0" for each partial image 601 is set. P0, the probability P1 of "class 1", and the probability P2 of "class 2" are stored. The determination result 524 is not limited to the probability for each class, and the class with the highest probability may be simply stored as the determination result.

Returning to FIG. 5, the cleaning condition 525 is a condition for determining whether or not cleaning of chips is necessary. As the cleaning condition 525, a condition that combines the position of each partial image and the probability for each class is set. For example, when the above-mentioned three classes are set, an efficient and high-performance cleaning determination algorithm is realized by evaluating the following cleaning conditions (1) to (4) in descending order. (1) Of P0, P1 and P2, the maximum value is P2: wash (2) P1 + P2 ≧ 99%: wash (3) the mesh area is on the table 12 and P1 + P2 ≧ P0. Yes: Wash (4) Not applicable to the above conditions (1) to (3): Do not wash For example, if the above-mentioned two most simplified classes are set, the class is regardless of the position of the mesh area. A cleaning condition may be set such that if it is 0 (without chips), it is not washed, and if it is class 1 (with chips), it is washed.

For example, the number of washings 526 includes, for example, the number of continuous washings, which is the number of times the inside of the machine is continuously washed, and the number of times of local continuous washing, which is the number of times each mesh area is continuously washed. Central Processing Unit). By executing the chip detection program 521, the arithmetic processing unit 501 executes the drive control unit 511, the image pickup control unit 512, the mesh division unit 513, the chip determination unit 514, the cleaning processing unit 515, the teaching processing unit 516, and the additional learning unit 517. And functions as a coolant control unit 518.

When the ATC camera is used as the camera 102, the drive control unit 511 controls an automatic tool changer (ATC) to change the tool to the ATC camera.

The image pickup control unit 512 outputs a shooting control signal to the camera 102 at a preset shooting timing, acquires an in-flight image 527, and stores it in the storage 502. The shooting timing is between steps, tool change, fixed time interval, timing when a predetermined machining amount is reached, and the like.

The mesh dividing unit 513 divides the in-flight image into a plurality of partial images. This makes it possible to simplify and improve the accuracy of chip detection.

The edges that make up the outer shape of the chips are relatively complex and irregular. This feature is not lost even when a large number of chips are piled up and folded, and individual shapes cannot be recognized. On the other hand, the work or jig that is the background of the chips generally does not have the irregularity of the chips. Therefore, the mesh dividing portion 513 divides into partial images of a determined size based on the sizes of the chips, the work, and the jig. This makes it easier to distinguish the edge shape of the work or jig from the chips, and erroneous detection is reduced. The mesh is not limited to a grid pattern, and may be a rhombus, a triangle, a honeycomb shape, or the like. Further, it is not necessary to divide the entire in-flight image into partial images, and only the area where chip detection is required may be divided.

The chip determination unit 514 executes preprocessing for improving the determination accuracy, such as noise removal processing, image size conversion processing, and the like, for each partial image. After that, the chip determination unit 514 determines the presence or absence of chips using the chip determination model 522 by executing the chip detection program 521 for each of the partial images, and saves the determination result 524.

The chip determination unit 514 uses, for example, a convolutional neural network (CNN), which is a model specialized for image classification among deep learning methods that are relatively resistant to environmental changes. However, the present invention is not limited to this, and other machine learning algorithms may be used.

For example, when using a machine learning algorithm that uses input data that is not an image, such as a support vector machine (SVM), a filter process or the like is executed on the partial image, and the shape of each partial image. Calculate the feature amount. Then, the class may be determined for each of the partial images by inputting the shape feature amount into the trained chip determination model. Further, a non-machine learning algorithm such as a deterministic algorithm may be used. For example, a partial image tends to be more complex with more chips and more high image frequency components. Therefore, the chip determination unit 514 may execute frequency analysis such as Fast Fourier Transform (FFT), compare the spectral statistic of each partial image with the determination parameter, and determine the class. .. However, the fast Fourier transform is highly environment-dependent, and if small screw holes other than chips, droplets of coolant, etc. are imaged in the partial image, high-frequency components increase and erroneous determination is likely to occur. For this reason, an environment-dependent component is eliminated by preparing an image of the entire area without chips in advance, dividing the difference image between this and the image of the entire area to be determined into a mesh, and performing a fast Fourier transform. You may.

The cleaning processing unit 515 determines whether or not cleaning of chips is necessary for each of the partial images, and drives the coolant pump 106 to perform cleaning. The cleaning processing unit 515 compares the stored cleaning condition 525 with the determination result 524 of each partial image. Then, it is determined that cleaning is necessary for the partial image satisfying the cleaning condition 525. When the cleaning processing unit 515 determines that cleaning is necessary, the cleaning processing unit 515 controls the coolant injection unit 101 to clean the chips. The cleaning processing unit 515 calculates the position of the mesh area (actual position in the machine tool 100) corresponding to the partial image determined to require cleaning, and discharges the cleaning liquid from the coolant injection unit 101 toward the mesh area. Inject.

The position of the mesh area can be specified, for example, by making an image memory (not shown) for storing an image of the entire area correspond to the coordinate system in the machine tool 100. The coolant injection unit 101 uses a programmable coolant nozzle 111 to inject coolant to all mesh areas determined to require cleaning while changing the injection direction.

The cleaning processing unit 515 updates the cleaning count 526 each time the cleaning operation is executed, and if the cleaning count 526 is less than a predetermined threshold value, cleaning is performed. On the other hand, when the number of washings 526 (at least one of the number of continuous washings and the number of local continuous washings) reaches a predetermined threshold value, a predetermined abnormality processing is executed.

The processing at the time of abnormality is not particularly limited as long as it is a processing for avoiding a situation in which the automatic cleaning operation is not terminated indefinitely. For example, the operation of the machine tool 100 may be stopped and an alarm may be notified. Alternatively, while recording and notifying that the threshold value has been reached, cleaning of the mesh area is interrupted for a "cleaning interruption period" set in advance by the user, and cleaning of other mesh areas and operation of the machine tool 100 are continued. May be good. The threshold value of the number of washings 526 is configured to be set by the user. When the number of cleanings 526 reaches a predetermined threshold value, there is a possibility that the injection method (correction amount C, injection pressure, diffusion level, moving method, etc.) selected for the chips at that position is incorrect. In this case, the injection method model 107 may be updated by trying other injection methods.

The teaching processing unit 516 reads the determination result 524 of each partial image from the storage 502 and displays it on the touch panel 503. Specifically, each partial image is displayed in a mode in which the determination result can be identified (for example, different colors for different determination results). For example, they are classified into the following four groups according to the values of the determination probability P0 of class 0 (without chips) and the determination probability P1 of class 1 (with chips), and are masked with different colors.

Group 1: P1 is 0 to 30% (P0 is 70 to 100%): Red Group 2: P1 is 30 to 50% (P0 is 50 to 70%): Blue Group 3: P1 is 50 to 70% (P0 is) 30 to 50%): Green Group 4: P1 is 70 to 100% (P0 is 0 to 30%): Yellow The user who confirms the judgment result from the image touch-selects any partial image that is erroneously judged. , You can enter the correct judgment result (different color). For example, the partial images color-coded into group 2 can be classified into group 1, and the partial images color-coded into group 3 can be classified into group 4. The teaching processing unit 516 corrects the determination result 524 according to the change input of the determination result by the user.

The additional learning unit 517 performs additional learning using the new teacher data. The additional learning unit 517 causes the chip determination model 522 to execute additional learning using the determination result (correct chip detection result by the user) corrected by the teaching processing unit 516 as teacher data. Then, the determination parameter 523 is updated according to the result of the additional learning. The additional learning unit 517 may upload teacher data to a learning server on the network for additional learning, and update the chip determination model 522 with the obtained determination parameters. After the teaching mode ends, the additional learning unit 517 starts additional learning in the background at a predetermined timing, generates and updates a judgment parameter that reflects the corrected judgment result, and updates the chip judgment model 522. do.

The coolant control unit 518 controls the coolant pump 106 via the control unit 105. Specifically, the coolant pump 106 is turned off in the observation mode in which the situation inside the machine is captured as it is and displayed on the touch panel 503, or in the chip detection mode in which the chip detection unit 103 detects chips.

As described above, according to the present embodiment, it is possible to effectively remove chips by learning the method of injecting the coolant.

[Other embodiments]
Although the invention of the present application has been described above with reference to the embodiments, the invention of the present application is not limited to the above-described embodiment. Various changes that can be understood by those skilled in the art can be made to the structure and details of the present invention within the technical scope of the present invention. Also included in the technical scope of the invention are systems or devices in any combination of the different features contained in each embodiment.

Further, the present invention may be applied to a system composed of a plurality of devices, or may be applied to a single device. Further, the present invention is also applicable when an information processing program that realizes the functions of the embodiment is supplied to a system or an apparatus and executed by a built-in processor. In order to realize the functions of the present invention on a computer, the technical scope of the present invention includes a program installed in the computer, a medium containing the program, a server for downloading the program, and a processor for executing the program. .. In particular, at least a non-transitory computer readable medium containing a program that causes a computer to execute the processing steps included in the above-described embodiment is included in the technical scope of the present invention.

The derivation unit 104 may derive the correction amount C according to the chip position 141 without using the injection method model 107. For example, the derivation unit 104 updates the storage unit that stores the chip position 141 and the correction amount C in association with each other, and the correction amount C stored in the storage unit according to the change of the chip position 141 after the coolant injection. It may have an update unit. Here, the correction amount C is the injection target position 142 and chips when the position where the vector indicating the direction of the coolant nozzle 111 collides with the plane parallel to the bottom surface of the machine tool (including the bottom surface itself) is set as the injection target position 142. It is a difference from the position 141 and is a plane distance. The injection target position 142 and the chip position 141 can be defined by positions on the XZ plane on the bottom surface of the machine tool 100. However, the definition of the correction amount C is not limited to this, and may be, for example, an angle formed by the direction of the coolant nozzle 111 and the direction connecting the position of the coolant nozzle 111 and the chip position. The lead-out unit 104 may lead out the correction amount C according to the distance between the position of the coolant nozzle 111 and the chip position 141. In the case of a configuration in which the coolant nozzle 111 itself can be moved, the lead-out unit 104 may derive the correction amount C in consideration of the moving speed of the coolant nozzle 111.

Claims (10)

A fluid injection unit equipped with a nozzle that can control the injection direction of the fluid,
An image pickup unit that captures the inside of a machine tool,
A chip detection unit that detects the chip position inside the machine tool based on the captured image acquired from the image pickup unit, and a chip detection unit.
A derivation unit that derives the correction amount of the injection target position of the fluid from the chip position according to the chip position.
A control unit that controls the injection direction of the fluid by the nozzle using the correction amount, and
Machine tool equipped with. The machine tool according to claim 1, wherein the correction amount is a difference between a position defining the direction of the nozzle and the chip position. The machine tool according to claim 2, wherein the position defining the direction of the nozzle and the chip position are defined as positions on a plane parallel to the bottom surface of the machine tool. The machine tool according to claim 1, wherein the correction amount is an angle formed by the direction of the nozzle and the direction connecting the position of the nozzle and the position of the chip. The derivation unit derives the correction amount according to the distance between the position of the nozzle and the chip position, or derives the correction amount according to the injection pressure of the fluid by the fluid injection unit. The machine tool according to any one of claims 1 to 4. (A) The derivation unit further derives the injection pressure of the fluid according to the chip position.
The control unit controls or controls the fluid injection unit so as to inject the fluid at the derived injection pressure.
(B) The derivation unit further derives a diffusion level defined by the shape or area of the region reached by the fluid, depending on the chip position.
The control unit controls or controls the fluid injection unit so as to inject the fluid at the derived diffusion level.
(C) The derivation unit further derives how to move the nozzle according to the chip position.
The machine tool according to any one of claims 1 to 4, wherein the control unit controls the fluid injection unit so as to move the nozzle by the derived movement method. The derivation unit uses the chip position using an injection direction model generated by machine learning the chip position, the injection direction of the fluid injected in the past, and the change in the chip position after the fluid injection as teacher data. The machine tool according to any one of claims 1 to 6, which derives the injection direction in which the fluid should be injected, which is predicted from the above. A fluid injection unit equipped with a nozzle that can control the fluid injection method,
An image pickup unit that captures the inside of a machine tool,
A chip detection unit that detects the chip position inside the machine tool based on the captured image acquired from the image pickup unit, and a chip detection unit.
Predicted from the chip position using an injection method model generated by machine learning the chip position, the injection method of the fluid injected in the past, and the change in the chip position after the fluid injection as teacher data. A derivation unit that derives the injection method to inject the fluid, and
Machine tool equipped with. The injection method model is a model generated by machine learning using the injection method of the fluid injected in the past, the distance between the nozzle and the chip position, and the change in the chip position after the fluid injection as training data. Item 8. The machine tool according to Item 8. A fluid injection unit equipped with a nozzle that can control the injection direction of the fluid,
An image pickup unit that captures the inside of a machine tool,
A chip detection unit that detects the chip position inside the machine tool based on the captured image acquired from the image pickup unit, and a chip detection unit.
It is a control method of a machine tool equipped with
A derivation step for deriving a correction amount from the chip position in the injection direction of the fluid according to the chip position, and a derivation step.
A control step that controls the direction of the nozzle using the correction amount, and
How to control machine tools, including.

Images